The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches installation instructions from GitHub and uses npx to install related skill packages from the inference-sh and belt-sh organizations.- [COMMAND_EXECUTION]: The skill utilizes the belt CLI tool via Bash to perform authentication (belt login), model discovery (belt app search), and image generation tasks (belt app run).- [INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill accepts user-defined text prompts and external image URLs as input for the belt app run command in SKILL.md.
Boundary markers: No explicit boundary markers or delimiters are present to separate user-provided input from the command structure.
Capability inventory: The skill possesses the capability to execute shell commands (belt) and interact with an external image generation service.
Sanitization: The instructions do not demonstrate any sanitization or validation of the input strings before they are passed to the CLI tool. This is a standard attack surface for AI-integrated tools and is documented here for awareness.

nano-banana